Quick Summary of Opening Session of “Visualisation and Presentation in Statistics”

From @Flygirltwo: @psychemedia What’s the event? #ouvpstats Looks really interesting, especially the identified need to tell good stories around the stats.
.@Flygirltwo event is “Visualisation and Presentation in Statistics” http://bit.ly/kQyBKW #ouvpstats

John Gower (http://statistics.open.ac.uk/People/j.c.gower) availability of enticing viz tools may pose risks/danger of “improper use” #ouvpstats
Can use diagrams to help yourself come to some understanding of something, but image not necessarily useful to others,
Tools okay when used properly, but they get much “improper use”/misused
Wrote a paper on this but can’t get it published….
Problems with public understanding of statistics

Michael Blastland ( http://bbc.in/mbl9eZ ) now on… #moreorless #ouvpstats “Numbers go up, and down…” http://bbc.in/jeXvHl
Encouraging people to play their way to an understanding BUT people don’t know what questions to ask/test http://bbc.in/ln3Czz #ouvpstats
Another stats game (hospital death chance calculator) but it didn’t really work.. Storytelling is missing http://bbc.in/jh95gM #ouvpstats
Need narrative arc to help people make sense of (their use of) interactive viz. Playing gapminder animation, with no Hans Rosling #ouvpstats
Imposition of narrative means experts can say “Hmm, I’m not sure I agree with that…” #ouvpstats
Office for National Statistics is great, but little use to public. ONSstats on Youtube helps address this http://bit.ly/jLHp6B #ouvpstats
ONSstats on youtube uses narrative to help explain the stats… http://bit.ly/jLHp6B #ouvpstats
Most people don’t necessarily get much out of charts with wiggly lines #ouvpstats
Need an understanding of the underlying issues in order to engage with interactive data visualisation meaningfully #ouvpstats
Does imposition of narrative destroy opportunities for open-ended exploration eg with interactive visualisations? #ouvpstats
Qn: many statisticians come from maths background; need interdisplinary team of eg storytellers and designers…? A: Yes #ouvpstats
Fundamental dilemma: how to throw topic open to curiosity whilst providing narrative way in? #ouvpstats

From @JackieCarter: Discussion at #ouvpstats music to my ears. See http://bit.ly/lL4iBE and links from it to lots of work on this at Mimas
From @PhilDRoberts: Following #ouvpstats from my desk, gutted not to be there, have added to the Archivist see http://bit.ly/jS7tX6

Next up: John Aldrich http://bit.ly/mQAF2Z #ouvpstats
Victorian statistics… 1838 “collection and comparison of Facts which illustrate the condition of mankind” #ouvpstats
History of Victorian stats graphics by Funkhouser “Historical Development of the Graphical Representation of Statistical Data” #ouvpstats
First diagram in Statistical Journal was a line diagram in 1841 by Daniel Griffin, Limerick Literary and Scientific Society #ouvpstats
Great figure of Victorian statistics – William Farr http://bit.ly/kJdXpm – on occasion, did pictures “for special reports” #ouvpstats
Farr – “temperature and mortality of London” time series eye candy http://bit.ly/mcfB2E #ouvpstats Also reported on Crimea War…
… as did Florence Nightingale #ouvpstats Nightingale’s rose etc http://bit.ly/lzEBLO “Statistical aesthetics lagging behind”
Farr’s and Nightingale’s diagrams recognised by reviewers as remarkable but never became part of standard fare of communications #ouvpstats
First economic diagram in Statistical Journal 1847 John Towne Danson (“journalist”) – stats since passing of 1844 Bank Act #ouvpstats
W. Stanley Jevon’s statistical atlas 1860 #ouvpstats
Mulhall’s dictionary of statistics – full of pictograms BUT “real statisticians don’t do diagrams” #ouvpstats
New breed of statistician in 1890s who did make more of diagrams #ouvpstats
Visual adventurousness of mid 1800s did not become routine but time series diagrams did become routine #ouvpstats

From @agdturner: @psychemedia I wonder if you know about the work of John Snow: http://en.wikipedia.org/wiki/John_Snow_(physician) #ouvstats
From @agdturner: @psychemedia And then developing on this Stan Openshaw: http://en.wikipedia.org/wiki/Stan_Openshaw #ouvstats
@agdturner yes – but talk was more on history of charts/graphs rather than geo #ouvpstats

Next up: David Spiegelhalter on visualising risk/uncertainty http://bit.ly/inRfP8 h/t to McCandless, Fry… #ouvpstats
Quick typography of probability… decimal, fraction, percentage, odds, line/bar/chart etc, static/dynamic #ouvpstats
Relevance of positive/negative framing: “1% of blah” vs “99% of blah #ouvpstats
“Icon arrays .. generally considered quite nice” [but…?!?] http://1.usa.gov/lrPHix #ouvpstats
Would Nightingale have used animation of flash had been around? Animation: http://understandinguncertainty.org/nightingale
Visual football stats/predictions kickoff.co.uk Dangers using circles (area/angles). #ouvpstats
Fox: rubbish pie chart http://bit.ly/lvVpwn #ouvpstats
“Icon arrays .. generally considered quite nice” [but…?!?] http://1.usa.gov/lrPHix #ouvpstats
Most psychology experiments “sadly small” (“and what is an f-test anyway…!?”) #ouvpstats
Simplified “x in y” language generally deemed bad practice (changing denominator then changing numerator) #ouvpstats
Don’t use probabilities in teaching probability trees… use eg a population of 100 or 1000 to develop the intuition #ouvpstats
Assessing risk/uncertainty in screening tests – animating false positives http://bit.ly/leNd83 #ouvpstats
“Cone of uncertainty” in eg hurricane forecasts #ouvpstats What’s wrong with these? eg http://journals.ametsoc.org/doi/abs/10.1175/BAMS-88-5-651

Author: Tony Hirst

I'm a Senior Lecturer at The Open University, with an interest in #opendata policy and practice, as well as general web tinkering...

%d bloggers like this: